Topic
AI analytics
"AI analytics" covers several different products that share a label. Asking a question in plain language, generating a written summary, forecasting a series, detecting an anomaly and acting on one are five separate capabilities with five separate failure modes.
One requirement holds across all of them: an answer has to be traceable to a query.
Start hereAI Analytics: Seven Categories, Seven Different QuestionsSeven kinds of AI analytics, each answering a different business question, and what the infrastructure behind them actually has to do. Written for teams embedding this in a product rather than running it internally.
Definitions
The categories of AI analytics.
- TermAugmented AnalyticsUnderstand how AI supports analytics workflow through automation, suggestion, and interpretation, and what to validate before shipping results to product users.
- ArticleAI-Powered Analytics: Reactive to Proactive DashboardsCustomer expectations for analytics are changing faster than most platforms can keep up. Here's what we're seeing.
- ArticlePredictive vs Traditional Analytics for SaaS TeamsTraditional analytics tells you what happened. Predictive analytics tells you what might happen next. But which approach belongs in your customer-facing dashboards?
Natural language
Natural-language querying and conversational BI.
- ArticleNatural Language Query Analytics: From Question to Governed AnswerA production NLQ system must clarify intent, enforce semantic and tenant scope, bound execution, and return evidence with every answer.
- ArticleConversational BI: When Dashboards Talk BackCustomers are asking follow-up questions your static dashboards can't answer. Here's where the industry is heading, and what you can ship today.
- ArticleChatGPT for Business Intelligence: What We're LearningCustomers want ChatGPT-like interactions with their data. Here's what that means for embedded analytics.
Automation
Automated insights and anomaly detection.
- ArticleAgentic Analytics: AI That Acts (Not Just Analyzes)AI agents that don't just surface insights, they decide and act. Here's what agentic analytics means for embedded platforms.
- ArticleAutomated Insights: From Detected Signal to Accountable ActionDesign automated insights as an evidence path: detect a signal, show its limits, bound the action, and measure the outcome.
- ArticleAI Dashboard Personalization: Beyond Role-Based AccessA practical framework for personalizing customer-facing dashboards without confusing behavior signals with intent or weakening tenant isolation.
Forecasting
Predictive analytics and machine-learning visualisation.
- ArticlePredictive Analytics Dashboards: Forecasts to ValueMoving beyond 'what happened' to 'what's next', and why embedding predictions changes everything.
- ArticleMachine Learning Visualization for AI InsightsThe best ML models are useless if users can't understand them. Here's how modern teams are making machine learning transparent through visualization.
- ArticleAI Data Visualization: Answers Need EvidenceEvaluate AI visualization through interpreted questions, governed semantics, authorization, query evidence, uncertainty, and safe follow-up context.
By the audience it reports to
The same topic, by industry.
- Sales and CRMPipeline and CRM reporting inside your product
- Operations and Supply ChainOperations and supply chain analytics, with per-number freshness
- FinanceFinance analytics, reconciled to the ledger
- Retail and EcommerceRetail and ecommerce analytics, with comparable store windows
- HealthcareHealthcare analytics, with source reconciliation and access by care relationship
- MarketingMarketing analytics for agencies, white-label per client
- HRHR analytics, scoped by reporting line